When rotating a still image or a moving image in an image rotation process, a memory for storing an image before rotation and a memory for storing an image after rotation are needed. This is because if a memory has a capacity for one image screen, when a pixel of the image before rotation is read out, rotated, and written in a predetermined position, information stored in the predetermined position is lost because of overwriting in which the pixel before reading is overwritten with the image after rotation by a write operation. For this reason, the memory has to have the capacity for two image screens to rotate an image of one screen.
For this reason, a method for rotation with a reduced memory capacity is disclosed in patent document 1.
An image rotation processing device of patent document 1 is composed of storage means for storing image data, rotation process means for generating one or more square blocks and performing a rotation process, format conversion means for performing conversion into either a data format that is rotatable or a data format that is not rotatable, read address generating means for generating a read address to the storage means, and write address generating means for generating the address in a predetermined order in a write access to the storage means, and operates as follows.
In the image rotation processing device, the blocks are read in the predetermined order, converted into the format that is rotatable by the format conversion means, and written into a temporary area of the storage means and the written image is read out in a read address generation order that is selected from an address generation pattern according to a rotation angle, converted into the format that is not rotatable by the format conversion means, and written into a block area after rotation to the storage means in the predetermined order.
[Patent document 1] Japanese Patent Application Laid-Open No. 2003-115993